Publishing Data Online

Lizzie Scholtus

Institut für Ur- und Frühgeschichte

Why publishing data?

Why publishing data online?

To share data
- Open Science
- FAIR Principles

Why publishing data online?

What is Open Science?

Free access to:
- Scientific publications
- Research data and metadata
- Hypothesis, methodologies, protocols, codes, formats and results
Participative science

Pillars of the Open Science
according to UNESCO’s 2021 Open Science recommendation.
Source: Wikipedia

Why publishing data online?

What is Open Science?

An organised community of various disciplines with convergent practices
Manifeste des Digital Humanities
A government strategy

French national plan for Open Science
Source: French Ministry of Higher Education and Research

Why publishing data online?

What are FAIR principles?

= a way of constructing, storing, presenting and publishing data in such a way as to make it more easily accessible, understood, exchangeable and reusable by the community. Introduced in 2014 in response to a growing desire to publish open data and to make the most of big data.

Easy to find >> file described by metadata (description of file content, authors, dates, etc.) and associated with a unique identifier to make it easy to find (DOI, HAL).
Accessible >> long-term storage of data and easy access to it access to metadata through open and standard protocols + access to data through clear licences (not necessarily open or free)
Interoperable >> data can communicate with each other = various types and formats of data can be cross-referenced give priority to open and widely shared languages and formats + permanent identifiers + alignment with repositories
Reusable >> must be facilitated by the use of common standards + clear, verified and well-described data >> directly reusable
clear and accessible user licence (not necessarily free).

Why publishing data online?

To share data
- Open Science
- FAIR Principles

To save data
- Digital data can die
  - Accident
  - Software evolution
- Version control

Where to publish data?

Institutional
- EU, Country, University

non-specialist

To deposit the data as it is
- Private (personal or institutional cloud)
- Public (Zenodo, Nakala, HAL, opendata@uni-keil…)
- Both (Gitlab, Github)

Private
- Free or not

Field specialised

To share the data with other datasets
- (Ariadne, ArkeoGIS, Landman…)

To create your database (Heurist)

Where to publish data?

Three platforms, Three uses

Git Platforms

Oldest (2008)
More people
Bought by Microsoft
Less ready to use
- Users need to pay to integrate elements themselves from third-party application

Open source
Completely free at the beginning (not any more)
continuous integration and DevOps workflows
Need a paid account for some functionalities

Gitlab

Description

Dedicated to versioning
Collaborative work
Code development

Inconvenients

works with command lines
Only a depot, not a real publication

Advantages

Branch system
Continuous integration
- Standalone pages
- Website
Deposit can be private or public
Licence attribution
institutional instance exist

Zenodo

Description

Dedicated to dataset deposit
Code can also be deposit
Created By CERN

Inconvenients

Not possible to modify a file
Fix deposit as for a publication

Advantages

DOI (Digital object identifier)
Enables easy citation
Various metadata
Version control
Deposit can be private or public
Licence attribution

ArkeoGIS

Description

Geographical informatic system online
specific to data of the past
Created by Strasbourg university

Inconvenients

Support only spatial data
Only datasets
Data need to be aligned to ArkeoGIS structure

Advantages

Data link with other datasets
Automatic language alignement
Automatic chronological alignement
LOD (Link Open Data)
DOI

ArkeoGIS

Description

Geographical informatic system online
specific to data of the past
Created by Strasbourg university

Inconvenients

Support only spatial data
Only datasets
Data need to be aligned to ArkeoGIS structure

Advantages

Data link with other datasets
Automatic language alignement
Automatic chronological alignement
LOD (Link Open Data)
DOI
Data under password login or completly open

What is LOD?

What is LOD? (Link Open Data)

Linking online data together using:
- Semantic web
- Uniform Resource Identifier (URI)
- Detailed metadata
- Resource Description Framework (RDF)

What is Semantic Web?

Structure and link information online

Using ontologies to build conceptual data modelling

Using normalised vocabularies
- Getty AAT for patrimonial data
- PACTOLS for patrimonial data
- GeoNames for places
- Periodo for chronologies
- VIAF for people

Enables data browsing

Conceptual model based on CIDOC CRM ontology

A new platform for humanities

What does this mean for you?

What does this mean to you?

The first step is to structure your data
Always precise your data with metadata
If possible use free format
Place your data in permanent backup structures (institutional or not) >> DMP
Specify licences and where possible use the most open ones

The French Data Management Environment

A government strategy

National plan for Open Science
Law for a numerical Republic
Aim:
- Research Data Management on a national level
- Dissemination and reproducibility